This gets interesting...
https://www.weibo.com/2424567755/H6MbusMFT
Was this a real leak of an existing/concept sensor or a quality fake of the idea to be employed three years later? In my opinion, an RGBW array (embedded inside QB pixel blocks) that no longer interferes so greatly with basic assumptions of demosaicing can be the way to go, at least as a compromise between lowlight capabilities of mono and color accuracy of standard arrays. I'm still looking up to the standard of standard high resolution Bayer arrays, but a less problematic RGBW can still prove itself worthy. As an unmentioned disadvantage tho, it'll be even more obvious on such a sensor to avoid the "full res" modes achieved through remosaicing, unless immensely corrected by pixel shifting, stacking algorithms.